Riassunto
The past two decades have witnessed a dramatic increase in the demand for analyses of health and public affairs program data. Governments at the federal and state levels have stimulated, if not mandated, much of this increased de- mand and spawned numerous not-for-profit and quasi-public organizations- research groups, lobbying organizations, consulting firms-who actively use analytic techniques. Program analysis is a broad term that encompasses activities such as pro- gram planning and prediction, program assessment, and program evaluation. Though purposes differ and techniques vary, the common denominator for all applications is the attempt to use quantitative techniques to provide compre- hensive and objective analyses. Program analysis in public health and public affairs shares another common feature: In many cases the data collected for these applications are categorical in nature-that is, discrete information rep- resented by categories. Whether a defendant in a criminal trial is convicted (yes, no), the number of times a person visits a community health center in a month (0,1,2,3, ...) , how a person feels about a proposed program (agree, disagree, not sure, no opinion)-these are examples of categorical data. This book describes a multivariate categorical data analysis technique- the weighted-least-squares (WLS) approach developed by Grizzle, Starmer, and Koch (GSK)-applied to program analysis in health and public affairs. It is written for in-service professionals who desire an introduction to applied cate- gorical data analysis and for preservice students who are studying quantitative methods.
Contenuti
I: Foundations.- 1. Multidimensional Contingency Tables.- 2. Guidelines for Analysis.- What Are Categorical Data?.- Is This a Categorical Data Problem?.- What Function Should Be Analyzed?.- What Sample Size Is Required?.- Are Zero Cells a Problem?.- How Many Analyses Should Be Performed?.- What Levels of Significance Should Be used?.- Summary.- 3. Contingency Table Analysis: The WLS Approach.- The Contingency Table.- Formation of Functions.- The Variance of a Function.- The Linear Model and Categorical Data.- An Example of WLS Analysis.- Additive Association.- Multiplicative Association.- Summary.- Exercises.- II: Simple Applications of the WLS Approach.- 4. One Response and Two Factor Variables.- Analyzing Sentences in Criminal Court.- Defining the Variables.- The Response Function.- Specifying the Model.- Estimating Parameters.- Evaluating the Model’s Fit.- Testing Hypotheses.- Interpreting the Model.- Summary.- Exercises.- Appendix: GENCAT Input.- 5. Interaction Among Factor Variables.- Adding Complexity.- Modeling Interaction.- A Modular Model.- Comparison of Models.- Summary.- Exercises.- Appendix: GENCAT Input.- 6. Mean Scores.- Measuring HMO Usage.- Formation of Means.- The Prediction Model.- Defining New Factor Variables.- A Reduced Model.- Using the Model.- Summary.- Exercises.- Appendix: GENCAT Input.- 7. Log-Linear Models.- Analyzing Air Pollution Effects.- The Data.- Defining the Variables.- A Rationale for the Logit.- The Logit Model.- Comparison of WLS and ML Results.- An Odds-Ratio Interpretation.- Lead—A Significant Factor?.- Summary.- Exercises.- Appendix: GENCAT Input.- III: Advanced Applications of the WLS Approach.- 8. Multiple Response Functions.- Examining Trust in Government.- Structure of a Multiresponse Table.- Mean Trust Scores.- Multiple Proportions.- Summary.- Exercises.- Appendix: GENCAT Input.- 9. Rank Correlation Methods.- Are People Consistent in Their Responses?.- A Rank Correlation Coefficient.- Computing Gamma Using GENCAT.- Examining Consistency.- Results: No Education Effect.- Summary.- Exercises.- Appendix: GENCAT Input.- 10. Rank Choice Analysis.- Choosing Among Competing Alternatives.- Discrete-Choice Data and Few Choices.- Discrete-Choice Data and Many Choices.- Continuous-Choice Data.- Summary.- Exercises.- Appendix: GENCAT Input.- 11. Follow-Up Life Table Analysis.- Predicting HMO Membership.- The Follow-Up Life Approach.- WLS and the Follow-Up Life Table.- Specifying the Model.- Age and Year of Enrollment.- Summary.- Exercises.- Appendix: GENCAT Input.- 12. Selected WLS Literature.- Paired Comparisons.- Observer Agreement.- Supplemented Margins.- Repeated Measures.- Partial Association.- Competing Risks.- Multiple-Record Systems.- Complex Sample Surveys.- Complex Functions.- Functional Asymptotic Regression Methodology (FARM).- Variable Selection.- Comparisons to Parametric Regression.- Philosophy of Data Analysis.- Appendixes.- Appendix A: Matrix Notation.- Matrix Definitions.- Matrix Arithmetic.- The Inverse Matrix.- System of Linear Equations—Scalar Presentation.- System of Linear Equations—Matrix Presentation.- Summary.- Exercises.- Appendix B: The Linear Model.- Traditional Approach to ANOVA.- Linear Model Approach to ANOVA.- Coding Methods.- Testing Hypotheses.- Two-Way ANOVA.- Interaction.- Summary.- Exercises.- Appendix C: Table of Chi-Square Values.- Appendix D: The GENCAT Computer Program.- An Overview of GENCAT.- Entering the Data to GENCAT.- Left-Hand Side of the Equation.- Right-Hand Side of the Equation.- Testing Individual Hypotheses.- Summary of Major Input to GENCAT.- GENCAT Input and Output for Chapter 4.- References.
Le informazioni nella sezione "Su questo libro" possono far riferimento a edizioni diverse di questo titolo.
